One conv3d head layer with 5x1x1 kernel followed by ReLU activation; ... with QuantStub/DeQuantStub and it is ready for Pytorch eager mode quantization.
確定! 回上一頁